Trapped
E. L. Cochran, F.
J. Adrian,
W
e are forced to forewarn the reader that the
juxtaposition of apparently contradictory
terms in our title is a result of evolution. When
organic chemists first began to understand molecular structure, the term radical came generally
to be used to refer to the parts of molecules. Thus,
in this context, a radical had only formal significance and referred to the molecular groupings
which would result if a particular bond in a
molecule were imagined to break. Somewhat later,
physical chemists found that when chemical bonds
actually do break, these molecular groups or radicals can have a transitory existence of their own.
These radicals, which were thought of only as
intermediates in chemical reactions, were called
free radicals. The study of free-radical kinetics soon
became a major field of chemistry, resulting in the
generation of much interest in the free radicals
themselves, their spectra, and their structures.
Techniques for obtaining free radicals in sufficiently high concentration for spectroscopic study
in the gas phase gradually evolved . These included
electric discharge and flash photolysis techniques.
More recently it was recognized that stable radical
concentrations of a few tenths of a percent could be
obtained by isolating the radicals from one another
T he authors wish to acknowledge the assista nce of S. N. Foner a nd
C . K . J en who, in 1954, originated the use of ES R tech n iques at APL
in the st udy of trap ped free radicals .
2
and V. A. Bowers
in inert matrices- thus the term trapped free radicals.
Molecules, with very few exceptions, have an
even number of electrons that are paired off in the
process of chemical- bond formation so that the
spin magnetic moments of the individual electrons
are cancelled. They are thus diamagnetic, generally
speaking. When a bond in a molecule is broken,
two free radicals are formed, each of which has an
odd number of electrons. Free radicals, therefore,
are characteristically paramagnetic.
The development of magnetic resonance spectroscopy has provided a convenient and powerful tool
for the study of magnetic materials. The branch of
magnetic resonance that applies particularly to free
radicals is called electron spin resonance (ESR) spectroscopy. The exploitation of this technique for the
study of trapped free radicals has resulted in much
progress in our understanding of the structure of
trapped radicals and of their interaction with their
environment.
The ESR Method
With ESR, as in all forms of absorption spectroscopy, we detect the species of interest by stimulating transitions from a lower to a higher energy
level and measuring the energy absorbed. The
energy levels utilized in ESR spectroscopy are the
Zeeman energy levels of the unpaired electron in
A PL T echnical D igest
Recent development of electron spin spectroscopy has
brought about greatly increased interest in trapped
free radicals, and, in many cases, has permitted a
m ore or less complete determination of their structure.
This work has also provided reliable guidelines for
predicting the structure of more complex radicals.
The radicals in the experiments discu ssed were formed
by means of elementary photochemical processes or by
simple addition reactions, providing knowledge of
basic chemical processes.
ADICA 5
an applied magnetic field. These Zeeman levels
result from the fact that the unpaired electron
possesses spin and a resultant magnetic moment.
The electron may be regarded as a small bar
magnet, restricted by quantum mechanical principles to be either parallel or antiparallel to an
external magnetic field. These two directions have
different energies, the difference (in frequency
units) being given by
v
g{3
= hHo,
(1)
where g (the ratio of the magnetic moment in
units of the Bohr magneton to the angular momentum in units of h/ 27r), {3 (the Bohr magneton)
and h (Planck's constant) are constants, and Ho is
the magnitude of the applied DC magnetic field. If
electromagnetic energy of frequency v is applied to
a suitable sample, the unpaired electrons will be
found to absorb energy at the value of the applied
magnetic field which satisfies Eq. (1). For these
studies the frequency used was ~9000 mc, corresponding to an applied magnetic field of ~3000
oersteds (oe).
Unpaired electrons in free radicals differ from
free electrons in that in the former the electron is
constrained to an electronic orbital, i.e., occupies
a well-defined region of space relative to the nuclear
framework of the radical. This means that the
January - February 1963
electron is intimately associated, and in a definite
manner, with one or more of the atomic nuclei in
the free radical. Since many of these atomic
nuclei have their own small magnetic moments
(about one thousandth as large as the magnetic
moment of the electron), their proximity to the
electron will alter slightly the magnetic field it
experiences. This magnetic interaction between
the electron and the nuclei of the radical is known
as the hyperfine interaction; its magnitude depends
on the relative orientation of the electronic and
nuclear magnetic moments. Because the nuclear
moment, like the electronic moment, is constrained
by quantum principles to certain preferred orientations, the magnetic hyperfine interaction will
split the ESR line into a discrete set of lines, one
for each permissible orientation of the nuclear
moment.
The allowed orientations of the nuclear moment
are described by the projection M I of the nuclear
spin angular momentum vector I along the
magnetic field direction; MI may assume the
values I, I - 1, ... , - I. Each atomic nucleus has
its characteristic value of I (although many possess
the same value). Values are integral or halfintegral in units of h/ 27r, and nuclei have been
found for which I is as great as 6. Thus for C12,
I = 0; for HI, I = Yz; for NI4, I = 1, etc. The
details of the hyperfine interaction are often very
3
complicated but, for a single nucleus, may be
described for one rather general case by adding an
appropriate term to Eq. (1):
v
= g(3 Ho + ~ MI
h
h
+ BMI
(3 cos
h
2
0-
I).
(2)
This is a good approximation only when A is substantially larger than B and when the electron
charge distribution is axially symmetric with regard to the nucleus in question- conditions which
are met in a surprisingly large number of cases.
As Eq. (2) indicates, the h yperfine interaction
with the nucleus may be divided into two components. The term AM II h is isotropic with respect
to the external field and arises because the electron
has a finite density at the position of the nucleus.
The magnitude of the parameter A depends on this
density as well as on the magnitude of the nuclear
moment. The component BMdh (3 cos 2 0 - 1),
depends on 0, the angle between the external
magnetic field and the symmetry axis of the
charge distribution of the unpaired electron. This
is the classical expression for the interaction of
two magnetic dipoles; the magnitude of B depends
on the magnitude of the nuclear moment and
decreases with the cube of the distance between the
electron and nucleus.
An additional assumption implicit in Eq. (2) is
that the g-factor is isotropic. It sometimes happens
that there is a small orbital contribution to the
total magnetic moment of the electron. This contriLI Q UID H e
---+~~~--~
LIQU ID N" --+~~
bution depends on the orientation of the radical
relative to the external magnetic field, and thus
makes the g-factor anisotropic. Additional terms
must be included in Eq . (2) to describe such cases
as this. Where the electron charge distribution is
axially symmetric with respect to the symmetry
axis of the free radical, and where for simplicity
we again consider hyperfine interaction with only
one nucleus, Eq. (2) becomes
v
= (ig n + igJ")(3Hol h
+ ~ MI +
(3)
[Hg n - gJ..)(3Ho + BM I J(3 cos Oe 2
1) l h.
Here, g II and g.L are the electronic g-factors for the
magnetic field, respectively parallel and perpendicular to the symmetry axis, and On is the angle
between the magnetic field and the symmetry axis
of the radical. Equation (3) is strictly valid only
when (g II - g.L) is much smaller than g II or g.L ,
which is the usual case for free radicals.
On the basis of Eq. (3) we can distinguish three
types of ESR spectra:
1. Fluid-media spectra, where (3 cos2 Oe - I)
is averaged to zero by the rapid tumbling of the
free radical. Here we observe sharp-line spectra
whose hyperfine structure is a result of A only;
2. Single-crystal spectra, where the free radicals are
all oriented in the same way by the crystal field, and
the spectrum is different for each orientation of the
crystal in the magnetic field; and
3. Spectra in polycrystalline solids and glasses, where
the free radicals are randomly oriented relative to
the magnetic field . Here, hyperfine structure is
due to A in Eq. (3), with the term in brackets
contributing only to line broadening.
In studying trapped radicals, one deals with the
latter two categories; at APL we have been almost
exclusively concerned with polycrystalline-type
spectra.
InstruInentation
GAS
]J
Fig. l - Liquid-helium cryostat shown in vertical
cross section.
4
In most of our work, trapped free radicals are
produced by the ultraviolet photolysis of a compound present in low concentration in an inert
matrix maintained at 4.2°K (the boiling point of
liquid helium at normal pressure). Figure 1 is a
diagram of the liquid-helium cryostat, which is a
modified version of a cryostat originally developed
at APL by W. H. Duerig and I. L. Mador for
optical spectroscopy. 1 The sample is collected on a
1 W. H . Duerig and I. L . Mador, " An Optical Cell for Use with Liquid
Helium," Rev. Sci . l nstr., 23 , Aug. 1952, 421-424 .
APL T echnical Digest
FREQUENCY STANDARD
KLYSTRON
I MAGN ET I
400POWER
AMPLIFIER
FIELD
MODULATION
CO ILS
TO MAGNET
POWER SU PPL Y
Fig. 2-Block diagralll of an ESR spectrollleter.
sapphire rod in direct thermal contact with liquid
helium. The rod, 2 mm in diameter and 50 mm
long, projects into the liquid - helium reservoir
through a vacuum-tight solder joint. The temperature of the sample is measured by means of a
thermocouple soldered to the rod.
The liquid-helium assembly can be raised or
lowered mechanically by means of a bellows arrangement. In the lowered position the sapphire
rod extends 16 mm into the rectangular cavity
which is resonant at a frequency of ~9000 mc.
The cavity is maintained at liquid-nitrogen temperature by thermal contact with the copper radiation
shield. The sample is deposited on the sapphire
rod through two entrance pipes terminating in
slits in planes 45 0 removed from the plane of the
drawing. During and/ or after deposition, the
samples are exposed to ultraviolet light from an
RF discharge in the tube shown. The lower wavelength limit of the light is selected by means of
appropriate window materials; lithium fluoride
(threshold = 1000 A), sapphire (1450 A), fused
quartz (1800 A), and Vycor 7910 (2400 A) are
used. In other experiments the window is replaced by a slit through which radicals may be
deposited from the gas phase in which they are
produced by electrical discharge or thermal dissociation in an oven. The effect of temperature on
the ESR of the sample is observed by removing the
liquid helium and allowing the sample to warm up
at a rate controlled by an adjustable heat leak.
The ESR spectrometer, shown in block diagram
in Fig. 2, is a standard bridge-type instrument.
The klystron tube, which is the microwave source,
January - Feb ruary 1963
is stabilized at the resonant frequency of the cavity
by frequency modulating the klystron at 5 kc over
a very narrow frequency region. If the klystron
frequency is not centered at the cavity resonance
frequency, this modulation produces a 5-kc error
signal that is fed back to the klystron in order to
correct its frequency to the cavity resonance. The
use of fixed tuned cavities makes it difficult to
sweep the frequency through an ESR absorption
line, so the normal procedure is to sweep the
magnetic field at fixed microwave frequency. As
may be seen from Eq. (2), these are equivalent
procedures. When the DC magnetic field satisfies
the resonance condition of Eq. (2), the sample will
absorb power from the microwave field.
This absorption is most easily detected by modulating the magnetic field strength very slightly at
400 cps, which causes a corresponding fluctuation
in the microwave power absorption. Since AC
signals are much more easily amplified and discriminated from noise than are DC signals, the use
of field modulation results in a large increase in
sensitivity over DC operation. Since the amplitude
of the magnetic field · modulation corresponds to
only a fraction of the ESR absorption line width,
the resulting signal is proportional to the slope of
the absorption line (Fig. 3). The use of 400 cps for
field modulation in our experiment is a compromise
between the higher signal-to-noise ratio obtainable
at higher frequencies (due to the noise characteristics of the crystal detector) and the decreased
eddy-current losses in the cavity walls at lower
frequencies. Somewhat better sensitivity may be
obtained at the expense of simplicity by using low-
5
I-
Z
L.u
a<:
a<:
::>
U
-'
<
t:;
>a<:
2
U
U
o
o
s:
-'
z<
S2
u
o
least is always obtained for all the nuclei which
interact significantly with the electron. By a careful
analysis of the line shapes for rigidly oriented
radicals in polycrystalline matrices, considerably
more data can be obtained. In this case the observed spectrum is the result of superposition of
spectra arising from all possible orientations of the
radical relative to the magnetic field. The resulting
line shape is illustrated for the axially symmetric
case (Eq. (3)) in Fig. 4. The two sharp peaks or
"lines" in the derivative of the broadened absorption line correspond to definite orientations of
the radical, namely OH = 0 and OH = 7r/ 2. Therefore, in certain cases where the various hyperfine
lines do not overlap appreciably, the broadened
line shapes may be analyzed to obtain the constants
in Eq. (3). When the magnetic intera.ctions are not
axially symmetric, the analysis is considerably more
complicated, though still tractable in favorable
cases. 2
MAGN ET IC FIELD H o
Fig. 3- ESR signal in the absence of field modulation (solid line) and with field modulation and
phase-sensitive detection (dotted line) .
frequency modulation together with superheterodyne detection.
SOllle Selected Results
Observation of ESR spectra of trapped free
radicals provides information of several types. First,
the unambiguous assignment of a previously unknown spectrum to a particular radical is in itself
useful for analytical purposes. Second, the spectrum
often provides knowledge about the physical environment of the radical. Thus, we have seen that
a freely-rotating polyatomic radical will have a
spectrum consisting of sharp lines, while a rigidly
oriented radical will be subject to line broadening
by g-factor or hyperfine anisotropies. In some cases,
it is possible to vary the temperature through the
transition region between these two extremes,
giving, at least potentially, the activation energy
for rotation in various matrices. For spherically
symmetric atomic species, A and g are usually
accurately known from atomic-beam studies. The
determination of these quantities for the trapped
atom therefore provides knowledge about the perturbing effects of the matrix.
A third type of information provided by the ESR
spectrum of a free radical relates to the wave
function of the unpaired electron and therefore to
the structure of the radical. Since A in Eq. (3) (or
in Eq. (2)) is proportional to if; (0) 2-the probability density of the unpaired electron at the
nucleus in question- this much information at
I
6
I
f-
o
Z
w
a<:
t:;
-'
<
z
S2
DERIVATIVE
OF ABSORPTION
a<:
V)
w
MAGN ETIC FIELD STRENGTH
Fig. 4 - Typical ESR line shape for a randoml y
oriented, non-rotating free radical having an anisotropic g-factor and an anisotropic hyperfine
interaction as described b y Eq. (3) . The strong positive peak in the derivati ve curve corresponds to the
magnetic field oriented perpendicular to the symmetry axis; the weaker negative peak corresponds
to the magnetic field oriented parallel to the
symmetry axis.
1
2 F . J . Adria n , E . L . Cochran , and V . A. Bowers, "ESR Spectru m a n d
Structure of t he Form yl Radical ," J . Chern. Phys ., 36, March 1962, 16611672.
APL T echnical Digest
Another class of information derives from the
fact that trapped radicals may be prepared by
photochemical processes which are themselves of
considerable interest. For instance, the products of
primary dissociations are observed directly, avoiding the difficulties of the usual inductive chemical
kinetic analyses, and simple radical reactions can
often be observed to occur in solid matrices at
4.2°K. From such observation we learn something
about the processes of solid-state chemistry. In the
following, we will describe some typical results that
have been obtained in these areas.
Hydrogen AtOlllS
Trapped hydrogen atoms are produced in the
photolysis at 4.2°K of many simple molecules such
as HI, NH a , PH 3 , SiH 4 , CH 4 , HCN, etc. Since
I = 7-2 for the hydrogen atom, we expect two
spectral lines. Because the electron charge distribution in this atom is spherically symmetric with
~---
I
r--
0)
Z
507.9 oe - - - - 1
.....- --1-1-- - 516.1 oe - - --1-\-- - - 1"
w
c:r:
r--
68.8 oe
V)
-----~
--.J
<:
z
MAGNETIC FIELD STRENGTH
~
I-Hf--- -
FR EE H A TOM ---t~......~
510 .2 oe
t-- 5 oe - - i
MAGN ETIC FIELD STRENGTH
Fig. 5- ESR spectrum of h ydrogen atoms produced
by photolysis of water in solid argon at 4.2°K.
respect to the nucleus (B = 0 in Eq. (2)), the
hyperfine interaction is described by the single
parameter A. The ESR spectrum for photolytically produced H atoms in a solid argon matrix is
shown in Fig. 5. We see that there are two different
kinds of H atom in this sample; one has a value of
A, 1.15 % greater than, and the other 0.46 % less
than, that for the free atom. 3 Significantly, only
the latter type of hydrogen atom is formed when a
gaseous mixture of atoms and molecules is condensed- a procedure that allows the sample more
time in which to equilibrate. A theoretical treatment 4 of the interaction of the atom with the matrix
3 E . L. Cochran , V. A. B owers, S. N. Foner, and C. K . Jen, "Multi ple
Trapping Sites for Hydrogen Atoms in Solid Argon," Phys. Rev. L etters,
2, Jan . 1959,43; a nd S. N . Foner, E. L. Cochran , V. A. Bowers, and C. K.
Jen, "Multiple Trapping Sites for Hydrogen Atoms in Rare Gas Matrices," J. Chem. Phys ., 32, April 1960, 963-971.
F. J. Adrian, "Matrix Effect on the Electron Spin Resonance Spectra
of Trapped Hydrogen Atoms," J . Chem. Phys. , 32, April 1960,972-981.
1
January - Februm'Y 1963
Fig. 6-ESR s pectrum of CH 3 produced by photolysis of CH31 in solid argon at 4.2°K.
indicates that the H atoms with the smaller value
of A are in substitutional sites in the face-centered
cubic rare -gas crystallites, while those with the
larger value of A are in a considerably more
cramped interstitial site.
Other photolytically produced atoms whose ESR
spectra have been studied in inert matrices include
deuterium, nitrogen, and phosphorus. In addition,
the alkali metal atoms have been studied by means
of a vapor-deposition technique. 5
Slllall Freely Rotating Radicals
Analyses of ESR spectra for a number of polyatomic radicals have been completed. Take for
example the spectrum shown in Fig. 6 for the
trapped methyl radical CH 3 which can be prepared by photolysis of methyl iodide 6 or other
suitable molecules in a solid argon matrix. We can
compute B as comparable to A for this radical,
and predict, therefore, that for a rigid and ran5 C . K. Jen, V. A. Bowers , E . L . Cochran, a nd S. N. Foner, "Electron
Spin Resonance of Alkali Atoms in Inert-Gas Matrices," Phys. Rev., 126,
June 1962, 1749- 1757.
6 E. L. Cochran, F . J . Adrian. a nd V. A. Bowers, "Anisotropic Hyperfine
Interactions in the ESR Spectra of Alkyl Radicals ," J . Chem. Phys ., 34,
April 1961, 1161- 1175.
7
domly oriented radical, the E SR lines will be
very broad. The simple sharp-line spectra can
therefore be explained only if the radical is freely
rotating at 4.2 °K. Similar results are obtained' for
NH2 and SiH 3 . In the case of CN we can actually
observe the transition from a rigidly oriented radical to a more-or-Iess freel y rotating radical as the
tempera ture is raised from 4.2 oK to 37 oK (Fig. 7). 7
The center line of this spectrum is narrow and
sharp at all temperatures because this line corresponds to M I = 0 (Eq. (3» and there is no g-factor
anisotropy. The weak line at the position of the
vertical arrow is of unknown origin but is presumably due to an impurity.
I
f-
0)
Z
w
e<:
fVl
--.J
«
z
~
e<:
Vl
W
Ethyl Radical
The ethyl radical is representative of a large
class of radicals which have only partial rotational
MAGNETIC FIELD STR ENGTH
Fig. 8- ESR spectrum of the ethyl radical produced
by photol ysis at 4.2°K of HI in s olid argon containing 9% eth ylene.
freedom. As may be seen from its structural formula,
H
'"
/
H ······C-G·····H,
/ {3
H
I
f-
0)
Z
w
e<:
fVl
--.J
«
z
~
e<:
Vl
W
t---
10 oe ~
(see text)
3274.5 oe
MAGN ETIC FIELD STRENGTH
Fig. 7-ESR spectrum of the CN radical produced
by photolysis of HCN in solid argon at 4.2°K.
E . L. Cochran, F. J . Adrian , a nd V. A. Bowers, "ES R D etection of t he
Cyanogen and Methy lene Imino F ree Radicals," J . Chem. Phys., 36 ,
April 1962, 1938-1942.
7
8
a'" H
there are two classes of H atoms in this radical:
those attached to the a carbon (a protons) and
those attached to the {3 carbon ({3 protons). The
spectrum for this radical (Fig. 8) shows a quartet
of relatively sharp lines due to the splitting by the
{3 protons. 6 This indicates that the unpaired electron
" sees" the {3 protons equally, i.e., that the radical is
free to execute internal rotation about the carboncarbon bond . Each of the lines of the quartet is
further split into a triplet by the a protons. The
outer lines of these subtriplets are broad because of
hyperfine anisotropy which cannot be averaged
out by the internal rotation alone. The center
lines of the triplets are somewhat sharper than the
side lines because these lines correspond to states
in which the two a-proton spins are opposed to
each other (antiparallel) . This, combined with the
internal rotation that makes the two a protons
spatially as well as chemically equivalent, permits
cancellation of the anisotropy due to the individual
protons. This spectrum shows, therefore, that the
ethyl radical is executing internal rotation and may
be undergoing overall rotation about the C- C
axis; it is not, however, free to execute end-over-
APL Technical Digest
end rotation. Similar results have been obtained for
other alkyl radicals. 5
Solid -State PhotocheIllistry
Photochemical processes in the solid state differ
in at least two important ways from those in liquid
or gaseous media. First, there are large "cage"
effects in the solid which tend to prevent the primary dissociation products from diffusing away
from one another. This promotes recombination
and drastically lowers the observed quantum yields.
These effects become more pronounced as the size
of the primary radicals increases. Thus, minor
dissociation modes in fluid media which involve,
say, the production of an H atom are sometimes
found to account for the major part of the decomposition in solid-state photolysis. Another consequence of the absence of diffusion in the solid is
that radical concentrations build up; and depending on their absorption spectra, they may
compete effectively for the actinic light. Thus,
products resulting from the successive absorption
of several quanta of light are observed. An example
of this is the photolysis of deuterated methanol,
CHaOD
~
CH20D
~
+
~CH20
'----v--'
H
+
~CHO
D
+
H,
for which the major radicals observed are H, D,
and CHO.2
PhotocheIllically Induced Addition
Reactions of H AtoIlls at 4.2°K
It has been found that when HI is photolyzed at
4.2°K in the presence of molecules containing
multiple bonds, addition reactions very often occur.
Since the activation energies for these processes are
not known, it is not clear whether they are " thermal" (with the H atom at 4.2°K) or "hot" (where
the H a tom possesses excess energy from the dissociation process). However, the 2537 A quantum
used in these experiments possesses 113 kcal of
energy while the HI bond strength is only 71 kcal.
The remaining 42 kcal of energy can only appear
as translational energy of the products if, as appears to be the case, the iodine atom is formed in
its ground electronic state. Conservation of momentum requires that 99 % of this go to the H
atom. Since the average translational energy of H
at 4.2 °K is only about 12 calories, it would appear
that the H atom would have to undergo many
collisions with the lattice before being reduced to
thermal energy. In a typical experiment, 10 % of
the matrix molecules possess multiple bonds, so
conditions would appear to be favorable for the
Jan uary - February 1963
occurrence of hot reactions. The following reactions
have been observed:
H
H
H
H
H
+ co
+ C2H 2
+ C2 H 4
+ CsH s
+
HCN
~
~
~
~
~
HCO
C2H a
C2 H 5
CS H 7
H 2 CN
(formyl )
(vinyl)
(ethyl)
(cyclohexadienyl)
(methylene imino)
There are clearly many other possibilities which
have not yet been tried. A photochemically induced
addition reaction is often the cleanest and most
direct way to prepare a radical of interest. Addition
reactions for methyl radical have not been observed, which is probably not surprising because
the factors that favor hot reactions in the case of
H atoms are distinctly less favorable for CH 3 •
Specifically, CH 3 has a larger number of degrees
of freedom over which to distribute excess energy;
its larger mass means that a larger fraction of the
excess translational energy will go to the iodine
atom; and finall y, the CH 3 will be able to transfer
translational energy more efficiently to the matrix
molecules. In addition, there may be steric factors
and activation energy considerations which could
help to account for the differing behavior of H
and CH 3 with regard to addition to multiple bonds .
These low-temperature reactions may provide
an important experimental approach to the elucidation of the detailed mechanism by which elementary reactions occur. As an example, when an
H atom is added to deuteroacetylene we obtain
the 1 ,2-dideuterovinyl radical,
H
+
DC=:==CD
~
HDC=CD.
The structure of the vinyl radical is not known but
should be planar by analogy with ethylene. Since
internal rotation about C-C multiple bonds cannot occur, there are thus two possible structures
for the 1 ,2-dideuterovinyl radical :
D
'"C=C/
'"
H/
(1 )
D
H
and
D
'"C=C/
'"
D/
(2)
Normally, since the energies of the two structures
are the same, we could expect that they would be
formed in equal amounts. In solid argon at 4.2°K,
however, we observe only one of these structures. A
comparison of the observed hyperfine splittings in
this radical, with theoretical estimates of these
quantities, indicates that structure (2) is formed.
This result is of interest because it is related to the
detailed mechanics of the manner in which the H
atom approaches the acetylene molecule and forms
the activated complex, which then relaxes to the
characteristic structure of the vinyl radical.
9